基于聚类算法及BP网络的零件族构造方法研究  

Research on Parts Family Formation Method Based on Clustering Algorithm and BP Neural Network

作　　者：刘冠玉 纪洪奎 冀万文 谢振平 LIU Guanyu;JI Hongkui;JI Wanwen;XIE Zhenping(Chengde College of Applied Technology,Chengde 067000,China)

机构地区：[1]承德应用技术职业学院,河北承德067000

出　　处：《新技术新工艺》2020年第3期56-60,共5页New Technology & New Process

摘　　要：为解决多品种小批量产品的零件成组设计与加工的问题,提出了基于编码的K-means聚类算法和有效性指标的已有零件成组,以及利用BP网络法实现新零件归族的方法来构造零件族的方法。建立零件聚类成组的数学模型,利用函数指标来检验零件成组的有效性并得出最佳聚类族数。利用神经网络算法来进行零件分组的BP网络训练,通过训练后的网络对新的零件进行仿真,从而实现新零件的匹配。通过实例证实该方法可准确构建相似件的零件族。To solve the problem of designing and manufacturing on the production of many varieties of small batch,the parts grouping method based on K-means clustering algorithm,clustering validity index and BP neural network method for new parts was proposed.Mathematical model of part clustering was built,then the effectiveness of parts group was tested by function index and optimal number of clusters group can be found.Furthermore,the grouped parts were used to train the BP neural network,then simulated new parts on network to find the match group.At last,a case study was also presented to verify the feasibility of this method to build parts family of similar parts.

关 键 词：成组技术 编码 聚类算法 零件族 有效性指标 BP网络 

分 类 号：TH163[机械工程—机械制造及自动化]